# AI Place Hub

> A community library of prompts, configuration files, tutorials and AI terminology. Written by the people who use them daily.

Everything here is written by people who use these tools daily. Each entry records where it was tested,
which part of it does the work, and what changes on other models. Content is licensed CC BY-SA 4.0.

Library size: 12 prompts, 5 system instructions, 6 configs, 3 cheatsheets, 2 workflows, 2 tools, 2 errors and fixes, 1 before and after, 1 local setup, 6 guides, 18 glossary terms.

## Structure

- [Prompts](https://aiplacehub.com/prompts): Single reusable prompts, filed by task and by the model they were tested on.
- [System instructions](https://aiplacehub.com/instructions): Standing instructions that tell the model who to be. Paste once into settings and every answer changes.
- [Configs](https://aiplacehub.com/configs): CLAUDE.md, AGENTS.md, Modelfiles, editor rules and parameter presets. Files, not personas.
- [Cheatsheets](https://aiplacehub.com/cheatsheets): Slash commands, keyboard shortcuts and formatting tricks, as a table you scan in ten seconds.
- [Workflows](https://aiplacehub.com/workflows): Multi-step procedures that chain several prompts into one finished job.
- [Tools](https://aiplacehub.com/tools): Apps, extensions, CLI tools and MCP servers, reviewed by people who run them.
- [Errors and fixes](https://aiplacehub.com/errors): The exact message you got, what causes it, and what actually fixes it.
- [Before and after](https://aiplacehub.com/before-after): The same task asked badly and asked well, with the output both produced.
- [Local setup](https://aiplacehub.com/local): Ollama, LM Studio, hardware and quantisation, without an API bill.
- [Guides](https://aiplacehub.com/guides): Longer explanations of why models behave the way they do.
- [Instruction builder](https://aiplacehub.com/builder): answer a few questions, get a system prompt to paste into settings
- [AI crawlers](https://aiplacehub.com/crawlers): user agents for GPTBot, ClaudeBot, Google-Extended and the rest, what each one does with your pages, and a robots.txt builder
- [Model rules](https://aiplacehub.com/models): context windows, output caps, prices per million tokens and typical refusals, with the date each row was verified
- [Glossary](https://aiplacehub.com/glossary): definitions of the terminology
- [AI systems](https://aiplacehub.com/ai): one page per model or platform
- [Collections](https://aiplacehub.com/collections): sets people assembled for one job and share as a single link

## API

JSON, no key needed: `https://aiplacehub.com/api/resources?type=prompt&platform=claude`
A single item with its full text: `https://aiplacehub.com/api/resources?type=prompt&slug={slug}`

## Model limits

- [GPT-5](https://aiplacehub.com/models/gpt-5): 400K context, $1.25 per 1M input, verified May 2026
- [GPT-5 mini](https://aiplacehub.com/models/gpt-5-mini): 400K context, $0.25 per 1M input, verified May 2026
- [Claude Opus](https://aiplacehub.com/models/claude-opus): 200K context, $15 per 1M input, verified May 2026
- [Claude Sonnet](https://aiplacehub.com/models/claude-sonnet): 200K context, $3 per 1M input, verified May 2026
- [Claude Haiku](https://aiplacehub.com/models/claude-haiku): 200K context, $0.8 per 1M input, verified May 2026
- [Gemini Pro](https://aiplacehub.com/models/gemini-pro): 1000K context, $1.25 per 1M input, verified May 2026
- [Gemini Flash](https://aiplacehub.com/models/gemini-flash): 1000K context, $0.3 per 1M input, verified May 2026
- [Grok 4](https://aiplacehub.com/models/grok-4): 256K context, $3 per 1M input, verified May 2026
- [Llama 3.3 70B](https://aiplacehub.com/models/llama-3-3-70b): 128K context, runs locally, verified May 2026
- [Llama 3.2 8B](https://aiplacehub.com/models/llama-3-2-8b): 128K context, runs locally, verified May 2026
- [Mistral Large](https://aiplacehub.com/models/mistral-large): 128K context, $2 per 1M input, verified May 2026
- [Mistral Small](https://aiplacehub.com/models/mistral-small): 128K context, $0.2 per 1M input, verified May 2026

## Instruction roles

- [Developer](https://aiplacehub.com/instructions/developer): Reads code before it changes it and says what it did not check.
- [Editor](https://aiplacehub.com/instructions/editor): Cuts, questions and leaves your voice alone.
- [Analyst](https://aiplacehub.com/instructions/analyst): Works from the numbers given and says when they are not enough.
- [Teacher](https://aiplacehub.com/instructions/teacher): Explains at the level you are at, not the level of the manual.
- [Copywriter](https://aiplacehub.com/instructions/copywriter): Writes for a reader, not for a word count.
- [Everyday assistant](https://aiplacehub.com/instructions/assistant): A general default that stops the padding.
- [Sysadmin](https://aiplacehub.com/instructions/sysadmin): Assumes production, asks before anything destructive.
- [Researcher](https://aiplacehub.com/instructions/researcher): Separates what a source says from what it concludes.

## AI systems covered

- [Universal](https://aiplacehub.com/ai/universal)
- [ChatGPT](https://aiplacehub.com/ai/chatgpt) - config file: AGENTS.md
- [Claude](https://aiplacehub.com/ai/claude) - config file: CLAUDE.md
- [Gemini](https://aiplacehub.com/ai/gemini) - config file: System instruction
- [Grok](https://aiplacehub.com/ai/grok)
- [Llama](https://aiplacehub.com/ai/llama) - config file: Modelfile
- [Mistral](https://aiplacehub.com/ai/mistral) - config file: Modelfile
- [Midjourney](https://aiplacehub.com/ai/midjourney)
- [Stable Diffusion](https://aiplacehub.com/ai/stable-diffusion)

## Most used right now

- [No-padding everyday default](https://aiplacehub.com/instructions/assistant/no-padding-everyday-default): The one to set if you set only one. Answer first, nothing before it, nothing after it, and no offer to help further.
- [Forcing an output format that holds](https://aiplacehub.com/cheatsheets/forcing-output-formats): The phrasings that actually constrain output, and the ones people use that do not. Tested by asking for the same thing twenty times and counting the…
- [ChatGPT commands and keyboard shortcuts](https://aiplacehub.com/cheatsheets/chatgpt-slash-commands-and-shortcuts): The commands and shortcuts that are not in the interface anywhere, collected in one table. Mostly keyboard, mostly undocumented.
- [Sysadmin who flags what cannot be undone](https://aiplacehub.com/instructions/sysadmin/sysadmin-who-flags-what-cannot-be-undone): For infrastructure work: gives the command, but names anything irreversible before it, and never assumes a test environment.
- [Senior engineer who reads before changing](https://aiplacehub.com/instructions/developer/senior-engineer-who-reads-before-changing): Turns the model into a colleague who answers the question you asked, touches only what you pointed at, and says out loud which parts it did not…
- [Editor who does not flatten your voice](https://aiplacehub.com/instructions/editor/editor-who-does-not-flatten-your-voice): Separates fixing what is wrong from rewriting what is fine, and only does the first unless you ask. Stops the slide into the same neutral register…
- [Teacher who checks where you are](https://aiplacehub.com/instructions/teacher/teacher-who-checks-where-you-are): Explains at your level rather than the manual's, by asking one question first and building from what you already said you know.
- [429 Too Many Requests / Rate limit reached](https://aiplacehub.com/errors/rate-limit-429-retry): You crossed a request or token limit for your tier. The fix is backoff with jitter and batching, not a shorter sleep, and the header tells you which…
- [This model's maximum context length is N tokens](https://aiplacehub.com/errors/context-length-exceeded): The request is longer than the model can hold, counting your prompt, the whole conversation history and the space reserved for the answer. Usually it…
- [Ollama commands worth knowing](https://aiplacehub.com/cheatsheets/ollama-commands-worth-knowing): The subset of the CLI you actually use, plus the two flags that fix the problems everybody hits.
- [Running your first local model](https://aiplacehub.com/local/running-your-first-local-model): Ollama on a normal laptop, from install to a model answering in about ten minutes, plus how to work out which size actually fits before you download…
- [LM Studio](https://aiplacehub.com/tools/lm-studio-for-comparing-local-models): A desktop app for downloading, running and comparing local models without a terminal. Best tool for deciding which quantisation is good enough before…
- [Ollama](https://aiplacehub.com/tools/ollama-for-running-models-locally): The simplest way to get a local model answering on your own machine. One command to install, one to pull a model, and an API on localhost that most…
- [Asking for a summary](https://aiplacehub.com/before-after/asking-for-a-summary): The same document, the same model. The weak version returns the article rearranged; the strong one returns something you can act on, because it says…
- [Turn a newsletter into posts for three networks](https://aiplacehub.com/workflows/turn-one-newsletter-into-three-social-posts): Takes one long email and produces a LinkedIn post, an X thread and a short video script, each written for how that platform actually reads, without…
- [Find dead code across a repository](https://aiplacehub.com/workflows/find-dead-code-across-a-repository): A four-step pass that maps the repo, lists exported symbols, checks each one for real callers and produces a deletion list you can actually act on…
- [Image prompt skeleton](https://aiplacehub.com/prompts/image-generation/image-prompt-skeleton): A fill-in structure for image models that keeps subject, environment, light, lens and style in the order the models weight them.
- [Giving a model a role and constraints](https://aiplacehub.com/guides/giving-a-model-a-role-and-constraints): How naming the audience, the format and the exclusions turns unpredictable answers into repeatable ones.
- [Debug from a stack trace](https://aiplacehub.com/prompts/coding/debug-from-a-stack-trace): Works from the trace and the relevant source to name the most likely cause, with the check that would confirm it.
- [Editor rules for a Django project](https://aiplacehub.com/configs/universal/cursorrules-for-a-django-project): A rules file that pins the Django and Python versions, the app layout and the things an assistant should never regenerate.

## Glossary

- [Agent](https://aiplacehub.com/glossary/agent): A model given tools and a goal, allowed to decide its own next steps in a loop rather than answering once.
- [Chain of thought](https://aiplacehub.com/glossary/chain-of-thought): Prompting a model to write out its intermediate steps before the final answer.
- [Context window](https://aiplacehub.com/glossary/context-window): The maximum amount of text, measured in tokens, that a model can take into account at once - your prompt and its answer together.
- [Embeddings](https://aiplacehub.com/glossary/embeddings): Numeric vectors representing text, where similar meanings land close together - the basis of semantic search.
- [Few-shot prompting](https://aiplacehub.com/glossary/few-shot-prompting): Showing the model two or three worked examples of the task inside the prompt, instead of only describing it.
- [Fine-tuning](https://aiplacehub.com/glossary/fine-tuning): Further training of an existing model on your own examples, to change its behaviour rather than its knowledge.
- [Hallucination](https://aiplacehub.com/glossary/hallucination): A confident, fluent answer that is simply wrong - an invented citation, function, statistic or fact.
- [Inference](https://aiplacehub.com/glossary/inference): Running a trained model to produce output - what happens every time you send a prompt.
- [Large language model](https://aiplacehub.com/glossary/llm): A model trained on large amounts of text to predict the next token, which in practice lets it write, summarise, translate and reason over text.
- [Multimodal](https://aiplacehub.com/glossary/multimodal): A model that accepts or produces more than one kind of input - text plus images, audio or video.
- [Prompt injection](https://aiplacehub.com/glossary/prompt-injection): An attack where instructions hidden in content the model reads get executed as if you had written them.
- [Quantisation](https://aiplacehub.com/glossary/quantisation): Storing a model with lower-precision numbers so it fits in less memory and runs on ordinary hardware.
- [Reasoning model](https://aiplacehub.com/glossary/reasoning-model): A model trained to spend extra computation working through a problem before answering.
- [Retrieval-augmented generation](https://aiplacehub.com/glossary/rag): A pattern where relevant documents are fetched first and pasted into the prompt, so the model answers from your data rather than its training.
- [System prompt](https://aiplacehub.com/glossary/system-prompt): Instructions given to the model outside the conversation, which apply to every turn rather than one message.
- [Temperature](https://aiplacehub.com/glossary/temperature): A sampling setting that controls how much randomness goes into choosing each next token.
- [Token](https://aiplacehub.com/glossary/token): The unit a model reads and writes - roughly a word fragment. Pricing, context limits and speed are all measured in tokens.
- [Vector database](https://aiplacehub.com/glossary/vector-database): A store built for finding the nearest vectors to a query vector quickly, across millions of items.

## Not here

No account areas, no private saves, no user email addresses. Pages under /admin, /api/like, /api/save,
/settings and /library are not public and are excluded from crawling.